According to experts like Dr. Anthony Fauci, we know we need 75-90% of the US population to be vaccinated to achieve herd immunity, contributing to a world without COVID-19. Those figures are similar in other countries, including Singapore and the UK. But even in large economies like the US, nearly 27% of Americans report they are vaccine hesitant.

A unique feature of this pandemic has been the torrent of both mis and disinformation, and the mainstream nature of it. According to a study conducted by Pew Research Center, in the US alone, a whopping 80% of respondents report seeing fake news about the pandemic, and that’s among those that are able to decipher it. With elected government officials openly discussing, and even promoting misinformation, it can be hard for even the most educated audiences to decipher what is real and what is fake.

Recent studies indicate that Facebook is the most used platform for sharing misinformation overall. Among those most likely to engage, ideology emerges as the biggest marker, but on both sides of the aisle. The people who share mis and disinformation tend to be either staunchly liberal or staunchly conservative. In relation to the COVID-19 vaccine, the vaccine resistant social media following has ballooned to at least 58 million followers globally on English language vaccine resistant channels alone. This phenomena is reported to be especially significant on private messenger apps like Facebook Messenger and WhatsApp, where 26% of respondents surveyed in a recent COVID-19 Consortium study say they believe COVID-19 misinformation when they receive it through these platforms. Because we are looking at stated survey data when arriving at that figure, it is likely to be even larger in reality.

Understanding the “why” behind vaccine mis and disinformation

The question communicators need to be asking themselves in order to drive vaccine acceptance and adoption is not if, but why masses of people are enthralled by vaccine resistant content. To answer this, our team at Edelman Data & Intelligence (DxI) analysed 40,000 US-based Twitter conversations between 6 December and 13 December, focusing specifically on the language used by online communities discussing COVID-19-related topics, such as the first time the Pfizer vaccine was administered on 8 December 2020.

We used natural language processing techniques, including multilabel text classification and topic modeling, and psycholinguistic approaches—pioneered by Professor James Pennebaker—to identify the array of psychological dimensions of people ranging from vaccine resistant to acceptant. In essence, we pierced through the rhetoric by analyzing the language people use to dive into the deeper motivations, driving the sentiments reflected in the language, rather than just focusing on the opinions being shared.

The vaccine resistors

Our analysis found that while the vaccine resistant cluster, at an aggregate level, is actually open to new ideas, they are far more emotional about vaccine-related topics than the pro-vaccine cluster. Notably, they use words associated with pain and anger towards vaccine-related topics, along with negative words commonly associated with depression, such as “hate” and “worthless”. Their language also exposed high levels of fear around the unknown effects of the vaccine, such as causing issues like auto-immune diseases, death, allergic reactions, and facial paralysis. Fear is a powerful driver as it is a perceived threat by a perceived enemy, and for this group, the vaccine is more fear-inducing than the virus itself.

Vaccine resistors also fear that the vaccine causes infertility, a fear predominantly held and discussed by men. Because the sample of Tweets reflects the perspectives of the general population, it is fair to say that this finding goes far beyond just men who might be actively looking to have children, exposing the deeply rooted emotional nature of the issue—this male group feels fertility is interconnected to their identity and sense of self. For this male group, infertility generates a feeling of sexual inadequacy and a lack of “manliness”, while both women and men tend to feel “defective” or “incompetent”.

Identifying the language patterns and core issues to address, underscores a clear opportunity for those looking to drive vaccine acceptance and adoption across this group. Because fear of infertility is a powerful driver and speaking directly to that fear will factual information only deepens it, an emotional and visceral messaging approach should be adopted to drive behaviour change.

That strategy could include a micro-targeted effort that reaches couples or parents that focuses on how they can best care for their family or start a family, and the ways in which the COVID-19 vaccine will enable that. Rather than demonizing them as “ignorant” or “science-haters,” communicators and public health officials should not fixate on the often-perceived abrasive rhetoric, and instead can use their messaging as a mechanism for expressing an understanding of their underlying motivations and core concerns. Ultimately, addressing these motivations and concerns is essential if communicators are to begin opening a dialogue with the vaccine resistors that fosters connection and understanding.

The vaccine acceptors

On the other end of the spectrum, the vaccine acceptant cluster at an aggregate level shows higher levels of conscientiousness, meaning they are more disciplined, controlled, and autonomous. Their orientation is towards work and career, along with their focus on science to form their vaccine views. This segment also displayed less emotion when discussing the vaccine. Instead, they focused on the evidence and scientific grounds for a vaccine but emphasized that the decision to take it is in fact a personal decision.

Even when discussing the four Bell’s Palsy cases, their responses concern the lack of scientific validity in dismissing the vaccine due to this incident. Messaging that is measured, evidence-driven, and emphasizes personal agency in vaccine adoption is important in continuing to sustain trust with this group. Additionally, the messenger is likely to be significant, and trusted science and medical leaders are likely to carry more weight.

Despite their accepting attitudes, this group’s language still indicates that autonomy in decision-making is important. Therefore, they might be even more empowered to take the vaccine so long as they feel it has been left as a personal choice they make themselves. The importance of making these autonomous vaccine acceptors feel that they have the freedom to make a personal decision to take the vaccine is especially notable for businesses and organisations that are considering making the vaccine mandatory. A mandatory vaccination policy could backfire and cause vaccine acceptant people to become vaccine resistant.

Resistors vs. acceptors

Even at a macroscopic level of analysis, comparing resistance to acceptance alone, enormous and sweeping messaging differences emerge that are critical to driving, building, and sustaining trust with each group of people (who have a multitude of diverse factions). The scientific, evidential, and personal choice orientation of the vaccine acceptant community would naturally fall on deaf ears amongst a highly emotional, uncertainty-ridden group of people that is deeply afraid of what they perceive to be potential vaccine side-effects, especially as it relates to issues surrounding fertility.

As a community of data scientists, communications strategists, and storytellers, we must acknowledge that this level of empathetic data analysis is tablestakes. For each audience group and associated sub-community, it is crucial to understand the motivations behind their language to develop content and identify messengers that cut through the noise and speak in ways that are emotionally resonant and relevant to each individual.

It starts with an unbiased quest to understand human beings, to dive below the surface, meeting their emotional and motivational states – a language that is often more powerful than facts. The result is a bridge built on trust.

A version of this post first appeared on World Economic Forum.